Chris Wikle

View Calendar
March 22, 2021 11:30 am - 12:30 pm

Adapting Efficient Reservoir Neural Methods for Statistical Long-Lead Forecasting of Ecological Processes

If it is so difficult to forecast the weather a few days from now, how can we forecast the state of weather-sensitive ecological systems one year out? The atmosphere is a chaotic dynamical system, and because of that, skillful weather forecasts are only possible out to about 7 to 10 days.  However, dynamical processes in the ocean operate on a much longer time scales, and many atmospheric processes depend crucially on the ocean as a forcing mechanism.  Some of these processes can be forced remotely, leading to so-called teleconnections.  This coupling between the slowly varying ocean and the faster varying atmosphere and associated processes, allows for the skillful prediction of some general properties of such systems many months in advance.  These long-lead forecasts are important for activities such as agricultural and wildlife management, energy production, and disaster planning, to name a few.  It has consistently been established that statistical models can provide some of the most skillful long-lead forecasts. However, it can be particularly challenging to specify parameterizations for nonlinear dynamical spatio-temporal forecast models that are simultaneously useful scientifically and efficient computationally, and that allow for proper uncertainty quantification.  In some cases, when such information is available, one can embed mechanistic information into multi-level (deep/hierarchical) models to facilitate parameter reduction and interpretability.  When such information is not available (such as with animal behavior) then alternative learning strategies such as deep neural models can be applied.  One challenge with such methods can be uncertainty quantification and the necessity of having very large data sets for model training.  Here I present an approach where we have integrated an alternative efficient “AI” spatio-temporal dynamical model, a deep echo state network, in a statistical framework to accommodate uncertainty quantification. This is applied to long-lead (one-year ahead) forecasting of settling patterns of Northern Pintails (Anas acuta) in the Prairie Pothole region of North America.

Related upcoming events

  • October 3, 2022 11:30 am - October 3, 2022 12:30 pm

    Methane Plume Mapping Over Offshore Oil and Natural Gas Platforms in the Gulf of Mexico

    Alana Ayasse

    Research Scientist

    Carbon Mapper

    University of Arizona


    Offshore oil and natural gas platforms are responsible for about 30% of global oil and natural gas production. Despite the large share of global production there is little known about the greenhouse gas emissions from these production facilities. Given the lack of direct measurements, studies that seek to understand the greenhouse gas contribution of offshore oil and gas platforms are incredibly important. The use of airborne remote sensing to map greenhouse gases from onshore oil and gas infrastructure has become a prominent method to quantify and attribute large individual emissions to their sources. However, until now, this method has not been used offshore due to the lack of consistent reflected radiance over water bodies.  In this talk I will present the results from a 2021 study where we used visible/infrared imaging spectrometer data collected over the Gulf of Mexico to map methane emissions from shallow water offshore oil and natural gas platforms. I will discuss the methods we employed to map methane in the offshore environment and how that differs from the onshore environment. I will show how remote sensing can efficiently observe offshore infrastructure, quantify methane emissions, and attribute those emissions to specific infrastructure types.


    Dr Alana Ayasse is a research scientist at Carbon Mapper and the University of Arizona. She earned her BA in Geography and Environmental Studies from UCLA and her PhD in Geography from UCSB. Her research focuses on improving remote sensing techniques to map methane and carbon dioxide plumes, understanding the role of satellites in a global carbon monitoring system, and using remote sensing data to further understand trends in carbon emissions.

  • October 10, 2022 11:30 am - October 10, 2022 12:30 pm


  • October 17, 2022 11:30 am - October 17, 2022 12:30 pm


  • October 24, 2022 11:30 am - October 24, 2022 12:30 pm


  • October 31, 2022 11:30 am - October 31, 2022 12:30 pm


  • November 7, 2022 11:30 am - November 7, 2022 12:30 pm


  • November 14, 2022 11:30 am - November 14, 2022 12:30 pm


  • November 21, 2022 11:30 am - November 21, 2022 12:30 pm


  • November 28, 2022 11:30 am - November 28, 2022 12:30 pm


  • December 5, 2022 11:30 am - December 5, 2022 12:30 pm